Updated pipeline.get_prediction_intervals() to add trend prediction interval information from STL decomposer

[christopherbunn]

Resolves #4060

[codecov]

Codecov Report

Merging #4093 (483a613) into main (6fd5daf) will increase coverage by 0.1%.
The diff coverage is 100.0%.

@@           Coverage Diff           @@
##            main   #4093     +/-   ##
=======================================
+ Coverage   99.7%   99.7%   +0.1%     
=======================================
  Files        349     349             
  Lines      37588   37644     +56     
=======================================
+ Hits       37469   37525     +56     
  Misses       119     119             

Impacted Files	Coverage Δ
...nents/transformers/preprocessing/stl_decomposer.py	100.0% <100.0%> (ø)
...valml/pipelines/time_series_regression_pipeline.py	100.0% <100.0%> (ø)
...nent_tests/decomposer_tests/test_stl_decomposer.py	100.0% <100.0%> (ø)
.../tests/pipeline_tests/test_time_series_pipeline.py	100.0% <100.0%> (ø)

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.

[eccabay]

Looks pretty good! Just a few comments about code conciseness and expected behavior.

[chukarsten]

Thanks for refactoring some of that code out of _project_trend(), definitely looks more concise. Left a few comments to address, but looking forward to getting this out!

[chukarsten]

Dumb question, and I'm guilty of doing this too, why not just have this be the default value in the function def?

[eccabay]

https://docs.python-guide.org/writing/gotchas/#mutable-default-arguments

[eccabay]

Of course, I suggested we use STLDecomposer.name, but now on the heels of Jeremy's work I feel like we should follow his lead and check for any(isinstance(c, STLDecomposer) for c in self.component_graph.component_instances.values() instead 😅 just to be safe

[eccabay]

Should we check here that there aren't any other cover values that got added somehow? i.e. when expected_coverage = [0.95] we only have the two columns?